Output_schema with LiteLlm

[Bar-Korkos]

Im trying to create an Agent with open-ai model, but i didn't manage to also use output_schema.

model = LiteLlm(model="openai/gpt-4o")
class PlatformRequestType(BaseModel):
    request_type: Literal["get_information_event", "analyze_dashboard_event", "other"] = Field(
        description="Type of user request being made"
    )
    confidence_score: float = Field(description="Confidence score between 0 and 1")
    description: str = Field(description="Cleaned description of the request")

root_agent = Agent(
    model=model,
    name="UIInterpreterAgent",
    instruction=PLATFORM_AGENT_PROMPT,
    # sub_agents=[],
    # tools=[],
    output_schema=PlatformRequestType,

)

this is the error I got
Invalid config for agent UIInterpreterAgent: output_schema cannot co-exist with agent transfer configurations. Setting disallow_transfer_to_parent=True, disallow_transfer_to_peers=True

[boyangsvl]

This is a current limitation of the model and it will be fixed in the future.

[fedorzh]

same problem here. and with the anthropic_llm as well

[stefanadelbert]

I'm getting this same problem when specifying model = "gemini-2.5-flash-preview-04-17", i.e. not using LiteLlm. The agent is a SequentialAgent and has sub-agents which themselves specify output_schema.

[zbloss]

Is there an ETA on a fix for this?

[sanketpatil06]

Hoping this issue is resolved, marking it as closed. Please feel free reach out for any further issue.